Remote-sensing applications for biomass estimation in mountain grasslands

Mountain grasslands are largely known for livestock production, but they also supply essential ecosystem services. In recent decades, they have experienced degradation due to natural influences and anthropogenic factors. As a result, monitoring activities become crucial for assessing their health and productivity. Aboveground biomass (AGB) is a key parameter for evaluating grasslands quality and as a proxy for ecosystem functioning. Its estimation has been studied for many years using remote sensing reflectance and vegetation indices (VIs). This research intends to evaluate the relationships between VIs and dry matter production, and how the use of high-resolution satellite data and geospatial analysis techniques can enhance the ability to monitor and manage mountain grasslands, providing valuable information for agricultural production, conservation efforts, and habitat management. The experiment was conducted in 2023 at Malga Carriola, a temporary agro- zootechnical site in Caltrano (Vicenza). Twenty-three botanical surveys recorded all plant species and their abundance, to draw a vegetation map of the grazing surface. The obtained matrix of species was subjected to hierarchical cluster analysis. Furthermore, a random selection of 130 points was performed, followed by plant height measures in July 2023 using a Rising Plate Meter. Grass was harvested in 1 m² units, dried, and weighed to assess the dry matter (DM). Dry Matter and vegetation height were then correlated with remote sensing data. Raster imagery from Google Earth Engine (Sentinel-2 data), focused on the NDVI Normalized Difference Vegetation Index (NDVI) and the Normalized Difference Thermal Index (NDTI) as key VIs for this research. Raster images were therefore imported into GIS for a full visualization of VIs’ trends, and for creating an attribute table for ANOVA to test the cluster analysis effect on DM and vegetation height. Despite previous research indicating a relationship between VIs and DM production, this research did not show any significant correlation. This study demonstrates the complexity of using NDVI and NDTI for AGB estimation in this area, suggesting that other factors can be involved in determining vegetation production and distribution. However, an accurate biomass estimation should include long-term monitoring and all plant growth stages. Moreover, integration of additional variables, testing multiple remote sensing platforms, and utilizing multiple VIs, can be important for defining accurate vegetation assessments, resulting in enhanced grassland protection activities.